understanding pragmatic markers.pdf

Understanding

PragmaticMarkers

A VARIATIONALPRAGMATIC APPROACH

KARIN AIJMER

UNDERSTA

NDINGPRAGMATIC

MARKERS

How do pragmatic markers vary according to the region, the socialsituation and the identity of the speaker? What can the study ofvariation pragmatics add to our knowledge of pragmatic markers?

Understanding Pragmatic Markers is inspired by recent ideas inpragmatics on variation in language use with regard to social, culturaland regional factors. Karin Aijmer argues that in addition to a generalanalysis of pragmatic markers we need to take into account theiroccurrence in different varieties of English, text types and activitytypes. Expanding the study of pragmatic markers in this way broadensthe area of study and results in a better knowledge of pragmaticmarkers in general.

By discussing the different theoretical frameworks, methodologiessuch as corpus linguistics and the debates around the categorisationof the functions of pragmatic markers, this accessible introductionbrings students and scholars in linguistics up-to-date with thisdynamic field of research.

Karin Aijmer is Professor Emeritus at the University of Gothenburg.

Cover image: @ CactuSoup/iStockphoto.

Cover design: McColmDesign.co.uk

www.euppublishing.com

ISBN 978-0-7486-3550-4

KARIN

AIJM

ER

Understanding Pragmatic Markers

AIJMER 9780748635498 PRINT.indd iAIJMER 9780748635498 PRINT.indd i 18/03/2013 14:2118/03/2013 14:21

AIJMER 9780748635498 PRINT.indd iiAIJMER 9780748635498 PRINT.indd ii 18/03/2013 14:2118/03/2013 14:21

UNDERSTANDING PRAGMATIC MARKERS

A Variational Pragmatic Approach

Karin Aijmer

AIJMER 9780748635498 PRINT.indd iiiAIJMER 9780748635498 PRINT.indd iii 18/03/2013 14:2118/03/2013 14:21

Karin Aijmer, 2013

Edinburgh University Press Ltd22 George Square, Edinburgh EH8 9LF

www.euppublishing.com

Typeset in 10/12 Ehrhardtby Servis Filmsetting Ltd, Stockport, Cheshire, andprinted and bound in Great Britain byCPI Group (UK) Ltd, Croydon CR0 4YY

A CIP record for this book is available from the British Library

ISBN 978 0 7486 3549 8 (hardback)ISBN 978 0 7486 3550 4 (paperback)ISBN 978 0 7486 3551 1 (webready PDF)ISBN 978 0 7486 8165 5 (epub)ISBN 978 0 7486 8164 8 (Amazon ebook)

The right of Karin Aijmerto be identifi ed as author of this workhas been asserted in accordance withthe Copyright, Designs and Patents Act 1988.

AIJMER 9780748635498 PRINT.indd ivAIJMER 9780748635498 PRINT.indd iv 18/03/2013 14:2118/03/2013 14:21

Contents

Acknowledgments viiiSymbols used in Discourse Transcriptions ix

1 Intr oduction 1 1.1 Introduction 1 1.2 Defi nition of Pragmatic Markers in This Work 4 1.2.1 Pragmatic Markers and Refl exivity 4 1.2.2 Pragmatic Markers as Contextualisation Cues 6 1.3 Methodology 8 1.4 Linguistic Theories Accounting for the Relationship between Pragmatic

Markers and Context 9 1.4.1 Integrative Theories 10 1.4.2 Relevance Theory 11 1.4.3 Pragmatic Markers and Meaning Potentials 12 1.5 Pragmatic Markers and the Context 13 1.6 Formal Features of Pragmatic Markers 16 1.7 Functional Features of Pragmatic Markers 17 1.8 Summary and Conclusion 17

2 The Pragmatic Marker Well 20 2.1 Introduction 20 2.2 Previous Studies of Well 21 2.3 Distribution of Well in the Corpus 25 2.4 Formal Properties of Well 27 2.5 Well and Collocation 29 2.6 Well and Meaning Potentials 29 2.7 Classifi cation of Well in This Work 30 2.8 Well and Coherence 32 2.8.1 Word- search and Self- repair 32 2.8.2 Well as a Turn- taking Device 34

AIJMER 9780748635498 PRINT.indd vAIJMER 9780748635498 PRINT.indd v 18/03/2013 14:2118/03/2013 14:21

vi UNDERSTANDING PRAGMATIC MARKERSvi CONTENTS

2.8.3 Transition According to an Agenda 35 2.8.4 Transition to a Quotation 36 2.9 Well and Involvement 37 2.9.1 Well and Agreement 37 2.9.2 Well and Disagreement 40 2.9.3 Well as a Feedback to Questions 40 2.10 Well and Politeness 42 2.11 Well in Private Dialogue 43 2.11.1 Well in Face- To- Face Conversation 43 2.11.2 Well in Telephone Conversation 49 2.12 Well in Public Dialogue 55 2.12.1 Well in Broadcast Discussion 56 2.12.2 Well in Cross- Examination 63 2.13 Well in Spontaneous Commentaries 66 2.14 Conclusion 71

3 In Fact and Actually A Class of Adversative Pragmatic Markers 74 3.1 Introduction 74 3.2 Previous Work 75 3.3 Distribution of In Fact and Actually over Text Types 75 3.4 In Fact 79 3.4.1 Introduction 79 3.4.2 Formal Factors 79 3.5 In Fact and Function 81 3.5.1 The Adversative In Fact 82 3.5.2 The Elaborative In Fact 85 3.5.3 In Fact as a Hedging Device 88 3.5.3 In Fact as a Softener in End Position 89 3.5.5 Summarising In Fact In Conversation 89 3.6 In Fact in Public Dialogue 90 3.6.1 In Fact in Legal Cross- Examinations 91 3.6.2 In Fact in Broadcast Discussion 95 3.7 In Fact in Monologues 98 3.7.1 In Fact in Demonstrations 98 3.7.2 In Fact in Unscripted Speeches 100 3.8 In Fact in Writing 102 3.9 Summarising In Fact 103 3.10 Actually 103 3.10.1 Introduction 103 3.10.2 Formal Factors 104 3.11 Actually and Function 107 3.11.1 Emphasising Reality 107 3.11.2 Explicit and Implicit Opposition 107 3.11.3 Hedging and Politeness 109 3.11.4 Novelty and Surprise 110 3.11.5 Emphasising the Speakers Position 111

AIJMER 9780748635498 PRINT.indd viAIJMER 9780748635498 PRINT.indd vi 18/03/2013 14:2118/03/2013 14:21

CONTENTS viiCONTENTS vii

3.11.6 Elaboration 112 3.12 Summarising Actually in Conversation 116 3.13 Actually in Public Dialogue 117 3.13.1 Actually in Classroom Lessons 117 3.13.2 Actually in Business Transactions 120 3.14 Actually in Monologues 122 3.14.1 Actually in Demonstrations 122 3.15 Actually in Writing 123 3.16 Summarising Actually 123 3.17 Comparison of In Fact and Actually 123

4 General Extenders 127 4.1 Introduction 127 4.2 Previous Work 128 4.3 Formal Structure of General Extenders 129 4.4 Data 131 4.5 Distribution of General Extenders across Varieties 132 4.6 Factors Accounting for Variability of General Extenders 137 4.6.1 Grammaticalisation 137 4.6.2 Function of And- and Or- Extenders 139 4.6.3 Shared Knowledge and Positive Politeness 140 4.6.4 Intensifi cation 143 4.6.5 Hedging and Negative Politeness 144 4.6.6 General Extenders and Fluency 145 4.7 Conclusion 145

5 Conclusion 148

References 150Author Index 158Subject Index 160

AIJMER 9780748635498 PRINT.indd viiAIJMER 9780748635498 PRINT.indd vii 18/03/2013 14:2118/03/2013 14:21

Acknowledgements

I am deeply grateful to many people who have read diff erent chapters of the book and pro-vided valuable comments. I especially want to thank Bengt Altenberg, Gisle Andersen, Ad Foolen, Jennifer Herriman, Joybrato Mukherjee and his students, Heike Pichler and Anna- Brita Stenstrm.

AIJMER 9780748635498 PRINT.indd viiiAIJMER 9780748635498 PRINT.indd viii 18/03/2013 14:2118/03/2013 14:21

Symbols used in Discourse Transcriptions

short pause.. short pause long pause. . . long pause= lengthening . . . incomplete words . . . untranscribed text[ ] speech overlap uncertain hearing(H) inhalation% glottal stop(TSK) type of vocal noise

AIJMER 9780748635498 PRINT.indd ixAIJMER 9780748635498 PRINT.indd ix 18/03/2013 14:2118/03/2013 14:21

AIJMER 9780748635498 PRINT.indd xAIJMER 9780748635498 PRINT.indd x 18/03/2013 14:2118/03/2013 14:21

1Introduction

Communication processes do not merely apply to society: they are indefi nitely varied as to form and meaning for the various types of personal relationships into which society resolves itself. Thus a fi xed type of conduct or a linguistic symbol has by no means necessarily the same communicative signifi cance within the confi nes of the family, among the members of an economic group, and in the nation at large. (Sapir [1931] 1951)

1.1 Introduction

A recent edited volume of articles on pragmatic markers describes the study of pragmatic markers as follows:

There are very many studies of discourse particles on the market, and by now it is almost impossible to fi nd ones way through the jungle of publications. For a newcomer to the fi eld, it is furthermore often very diffi cult to fi nd the bits and pieces that constitute an original model of the meanings and functions of discourse particles. Moreover, the studies available so far are hardly comparable: the approaches vary with respect to very many diff erent aspects: the language(s) under consideration, the items taken into account, the terminology used, the functions considered, the problems focussed on, and the methodologies employed. (Fischer 2006: 1)

Research on pragmatic markers has avalanched in recent years and pragmatic markers have been promoted to a major area in pragmatics as shown by the large number of approaches devoted to the topic. The approaches are synchronic and diachronic, formal and informal; approaches building on text- linguistic models; models of general cognitive processing or interactively relevant domains of discourse; as well as approaches concen-trating on syntactic, semantic, pragmatic, or prosodic aspects (Fischer 2006: 1). Notwithstanding the recent spate of research on pragmatic markers from diff erent perspectives some challenges remain. The approach in this book is inspired by recent

AIJMER 9780748635498 PRINT.indd 1AIJMER 9780748635498 PRINT.indd 1 18/03/2013 14:2118/03/2013 14:21

2 UNDERSTANDING PRAGMATIC MARKERS

ideas in pragmatics on variation in language use which can be explained by factors such as text type, activity type, speakers, region, degree of formality, etc. The time now seems ripe for a book on pragmatic markers paying attention to their variation with regard to social, cultural and regional factors. In an earlier study of prag-matic markers, I wrote that the sociolinguistic analysis does not have a prominent place (Aijmer 2002: 213). However, it was pointed out (ibid. 220) that we need to bring in sociolinguistic factors such as the type of speaker and, in particular, text type in order to describe how they diff er from each other. The hypothesis in this book is that in addition to a general analysis of pragmatic markers we need to take into account their occurrence in diff erent varieties (of English), text types and activity types. The aim of this work is accordingly to study some selected pragmatic markers and their distribution across diff erent text types, dialogues or mono-logues, situations (activity types), varieties of English and across time. However, there are many contextual variables and the present study only presents a modest beginning in the study of pragmatic markers in diff erent contexts. A major inspiration for studying pragmatic markers in variable contexts is the emer-gence of a new discipline: variational pragmatics (Barron and Schneider 2009: 426) and new ideas about factors causing variation. Variational pragmatics has the goal of examining pragmatic variation across geographical and social varieties of language, and determining the impact of such factors as region, social class, gender, age and ethnicity on communicative language use (Schneider and Barron 2008: 1). Variational pragmatics can be said to go hand in hand with a critical attitude to classic variational studies of socio-linguistic variables since the methods used in this tradition may be less suitable to study discourse variation. The aim is to go beyond a concern with sociolinguistic categories such as age and gender by introducing a pragmatic level on which language is studied as speech acts or in terms of turn- taking and topic organisation. Thus variational pragmatics highlights both the failure of sociolinguistics to address the pragmatic level of language to any systematic extent and the failure of pragmatics to address variation due to macro- social variables such as region, socio- economic status, ethnic identity, gender and age (Barron and Schneider 2009: 436). The majority of studies in variational pragmatics have so far mainly dealt with regional and social diff erences in the realisation of speech acts. By now we have a number of para-digm studies showing how speech acts such as requests vary across diff erent regional and social varieties. As a result we know quite a lot about the issues, the social factors involved and the methodologies used to study the variation of speech acts. But there is also a need for variational studies of pragmatic markers. Like speech acts, pragmatic markers are used variably depending on the region, the social situation and the identity of the speaker. Expanding the study of pragmatic markers to look at their variation depending on such contextual factors broadens the area of study and can also be expected to result in a better knowledge of pragmatic markers in general. Pragmatic markers have been studied earlier from a sociolinguistic point of view. However, the focus in earlier studies has been on the infl uence of sociolinguistic variables such as age, gender and class on the use of pragmatic markers. As a result we have some knowledge about their dependence on the users. For example, in a number of articles Janet Holmes has studied the form and functions of pragmatic markers and their use by men and women in New Zealand English (1986, 1988a, 1988b, 1990) and Andersen


INTRODUCTION 3

(2001) studied like and innit in youngspeak linking the sociolinguistic facts to gram-maticalisation. However we can also expect pragmatic markers to be used diff erently depending on text type and social activity. We may now be watching a change of focus. According to Pichler, the usage of discourse- pragmatic features is strongly constrained by the interactional and situational context of their occurrence to the extent even that these factors may outrank the eff ect of social factors on discourse variability. Among the wide array of factors which may be assumed to have an eff ect on variation and change Pichler mentions, for example, discourse type and activity context, topic, purpose of and attitudes to the interaction, speaker roles and relationship and the communicative channel. These features can be refl ected in frequencies and in diff erent functions (Pichler 2010: 584). Another approach is to broaden the study of pragmatic markers to diff erent regional varieties. Pragmatic markers have been shown to be used diff erently in British and American English although some of the diff erences are quantitative only (cf. Aijmer 2002). It is also the case that new pragmatic markers (or uses of pragmatic markers) travel quickly to other varieties. Like and be like (the latter referred to as an innovation in American English) are among the fastest- spreading constructions in English today (Mair 2009: 22). Be like is for instance spreading in American English and is reported as an innovation, for example, in Australian and Canadian English (cf. Mair 2009: 22 for references to ongoing research). In this work I will be less concerned with factors such as age, gender and class and focus instead on the importance of text type (used here with reference to both speech and writing), the social situation (chat, dispute, etc.) and the regional variety of English. We cannot discuss pragmatic markers in diff erent text types and varieties of English without considering where we are now in the area of pragmatic research. Recently much has happened in the discussion of the theoretical framework and the categorisation of the functions of pragmatic markers (cf. Innes 2010: 96; Fischer 2006). However there is little agreement on basic issues such as the defi nition of pragmatic markers, terminology, and how many meanings they can have. As Pons Bordera (2006: 94) describes the present situation, pragmatic marker research is a melting pot of problems and perspectives. (See Fischer 2006 for an extensive discussion of these larger issues and controversies.) By studying the formal and functional variability of pragmatic markers in diff erent text types and situations, we can still hope to contribute to the discussion of larger issues such as the infl uence of contextual variables on the function and meaning of pragmatic markers and whether they have a single meaning. The studies of the individual pragmatic markers are corpus- based and made possible by the existence of spoken corpora which include several varieties. A number of diff erent corpora will be used to examine how their frequencies, form and function vary across diff erent regional varieties, text types, activity types, etc. (see Section 1.3). The present chapter is a general chapter where I discuss some basic assumptions underlying the empirical case studies. The defi nition of pragmatic markers in this work is discussed in Section 1.2. Pragmatic markers are described as refl exive or metalinguistic indicators (Section 1.2.1) and as contextualisation cues (Section 1.2.2). Section 1.3 deals with the corpus- linguistic methodology. Section 1.4 discusses the description of the rela-tionship between pragmatic markers and context in some linguistic theories focusing on integrative theories (1.4.1), relevance theory (1.4.2) and especially the theory of meaning



potentials (1.4.3). The contextual parameters and the importance of stance are discussed in Section 1.5 and the formal features of pragmatic markers in Section 1.6. The functions of pragmatic markers are dealt with in Section 1.7. Section 1.8 summarises the discussion in this chapter. The bulk of this book consists of case studies of a few selected pragmatic markers. The pragmatic marker well occupies a central position. This is not surprising since it is one of the most frequent markers. It has been studied extensively in earlier work. However these works have focused on large issues such as whether well is a pragmatic marker, what core meaning it has, whether it is homonymous or polysemous, etc. Moreover well has been studied primarily in conversation. As a result, we know very little about its func-tions in other contexts although we are beginning to see some work which problematises the fact that well is used in new or diff erent ways when we move away from conversation (see also Chapter 2). The pragmatic markers actually and in fact discussed in Chapter 3 seem to have similar meanings and can sometimes be interchangeable. However, a more detailed study of their usage domains shows that there are both similarities and diff er-ences between them and that it is important to consider their uses in diff erent situations and text types. Pragmatic markers such as and that sort of thing and or something vary both with regard to age, gender and with regard to diff erent dialects as has been shown in earlier work. In Chapter 4 their usage is studied in diff erent national varieties of English.

1.2 Defi nition of pragmatic markers in this work

The terms contextualisation cue and metalinguistic indicator have been used in sociolin-guistic and anthropological work to describe pragmatic markers with reference to their functions in the communication situation and their relation to the context. These terms correspond well to how I understand the general function of a pragmatic marker as will be obvious from the discussion in the following sub- sections. The aim of this and the following sections is to defi ne the role of pragmatic markers within a general pragmatic theory focusing on the language user and the relationship between meaning and context. Pragmatic markers are surface phenomena. On a deeper level they are refl exive i.e. they mirror the speakers mental processes as envisaged in the fabric of talk- in- interaction commenting on what goes on in the speakers mind (Redeker 2006).

1.2.1 Pragmatic markers and refl exivity

Refl exivity is manifested as the speakers awareness of the linguistic choices made both with regard to what to say and how to say it (Verschueren 1999: 187). Speakers have access to their own speech production and closely attend to what is going on; they are metalinguistically aware of what type of interaction they are involved in, if something goes wrong in the process, and what their attitudes are. The speakers cognitive processes are hidden to observation. However, pragmatic markers (and other devices) can emerge as overt indicators of (or windows on) ongoing metalinguistic activity in the speakers mind.

Speech is permeated by refl exive activity as speakers remark on language, report utterances, index and describe aspects of the speech event. [. . .] This refl exivity


INTRODUCTION 5

is so pervasive and essential that we can say that language is, by nature, funda-mentally refl exive. (Lucy 1993:11; also quoted in Aijmer et al. 2006: 106)

Pragmatic markers function as indicators of metapragmatic awareness along with other features such as prosody, hesitation, pausing: a wide range of indicators of metapragmatic awareness passes under the variable and often overlapping labels discourse markers, dis-course particles or pragmatic particles (Verschueren 1999: 189). This property accounts, for example, for the ability of pragmatic markers to refl ect on and organise the discourse, for example to make it more coherent on the local and global level. Well for instance remarks on the planning going on in the speakers mind or accompanies processes such as reformulation or revision. It can be described as a refl exive utterance signal which can convey conversational uptake (feedback on a preceding utterance) or project a new utterance. It can also change the context (e.g. signalling a new stage of a social activity). Their ability to project a new stage in the discourse (a new activity, speech act, or text) is an important aspect of metalinguistic indicators. They therefore have a crucial role in controlling and changing the progress of the discourse. We can distinguish two ways in which a pragmatic marker (or another indexical sign) can project an element of the context following Silverstein (1992). Silverstein distinguishes between metalinguistic indicators which presuppose something about their context- of- occurrence and those which alter some aspect of the existing linguistic context.

Any indexical sign can index the context in two ways which have to do with the appropriateness of language or its eff ectiveness: Any indexical sign form in occurring . . . hovers between two contractible relationships to its contextual surround: the signal form as occurring either PRESUPPOSES (hence, indexes) something about its context- of- occurrence, or ENTAILS (CREATES) (and hence indexes) something about its context- of- occurrence, these co- present dimensions of indexicality being sometimes seen as essential properties of the signs themselves, appropriateness- to- context- of- occurrence and eff ectiveness- in- context- of- occurrence. (Silverstein 1992: 36; quoted from Haviland 1996: 280)

Presupposing depends upon aspects of the context which exist independent of speech while creative forms alter it. What is interesting in this connection is that pragmatic markers have the defi ning property that they can be creative: they index the context by altering an existing context and creating a new (linguistic or social) context. What the new (social) context is depends for example on the activity (the context entailed by the pragmatic marker is diff erent in a debate and in a classroom lesson). Speakers may thus use pragmatic markers to update the interlocutor about an upcoming shift to a diff erent topic, a new stage in the debate or the end of the classroom lesson. In addition to the function of monitoring the speakers progression through the dis-course any indexical sign can have a rhetorical or dialogical function in the interaction (cf. Fischer 2000: 283). This is the case when speakers use a pragmatic marker with a rhetorical function to take up stances either agreeing or disagreeing with what is said and with the hearer (cf. Section 1.5).



1.2.2 Pragmatic markers as contextualisation cues

Pragmatic markers typically mark off segments in the discourse thus helping the hearer to understand how the stream of talk is organised. However, positioning within the discourse is not suffi cient to explain how the propositional content is to be interpreted, in particular if we analyse the uses of pragmatic markers in diff erent text types and situa-tions. According to Gumperz (1996: 379):

[c]onversational analysts seek to account for the workings of speech exchanges through turn- by- turn examination of their sequential ordering (Schegloff 1986). Yet while interpretation always depends on how acts are positioned within the stream of talk, positioning alone is not enough. Even with relatively simple utterances, propositional content can only be assessed . . ., with reference to shared frames or common ground.

When we examine pragmatic markers in other text types than conversation they can be used in a number of specialised ways which can only be explained with reference to the characteristic features of the text type or the situation where they are found. Pragmatic markers are for example used to signal the transition to a new topic, activity, argument, stage in a narrative, a new speaker in a debate, the drawing- to- a- close of a telephone conversation, etc. Studying pragmatic markers in institutionalised discourse draws atten-tion to the fact that we need to consider a more abstract context than positioning or the immediate linguistic context, namely the interactants shared knowledge of the speech event. Such shared knowledge is especially important if the roles of the speakers and the organisation of the discourse are fi xed to some extent. In De Finas (1997) study of Spanish bien in the classroom the interpretation is made on the basis of a teaching frame where the teacher has a privileged role and bien is inserted in slots marking the transition from one classroom activity to another. De Fina (1997) described bien (English well) as a contextualisation cue quoting Gumperz:

A contextualization cue is one of a cluster of indexical signs . . . produced in the act of speaking that jointly index, that is, invoke, a frame of interpretation for the rest of the linguistic content of the utterance. (Gumperz 1996: 379)

Gumperz restricted himself to cues such as prosody, gestures or shifts (code switches or style shifts) which are generally less salient than pragmatic markers. However, prag-matic markers are equally important as contextualisation cues. They can, for instance, signal a change in the social context (a change of frames or an activity within a frame) and enable hearers to infer the speakers strategies and intentions on a moment- to- moment basis in the interaction (what is happening now, who is doing what): As metapragmatic signs, contextualization cues represent speakers ways of signaling and providing infor-mation to interlocutors and audience about how language is being used at any point in the ongoing stream of talk (Gumperz 1996: 366). The notion frame refers to the way an activity is intended by the participants and the


INTRODUCTION 7

way their roles are understood within it (Tannen and Wallat 1987; quoted from De Fina 1997: 347). For example, if well occurs in the context of a narrative we can assume that it marks diff erent stages of the narrative (cf. Norrick 2001). Well in this case functions as an instruction or cue to the hearer to interpret what follows in relation to common or shared knowledge about how a narrative is built up. Compare Gumperz: Co- participants who perceive and respond to the shift are then led to resort to their background knowledge and by an inferential process akin to Gricean implicature, to derive contextual presup-positions in terms of which the signs can be understood (1996: 37980). Thus the description of well as a contextualisation cue explains why a conversational move introduced by well can be interpreted diff erently depending on the expecta-tions associated with the particular speech activity where it is used. For example, in classroom lessons well is used to facilitate tasks such as indicating the opening of a lesson or showing whether the conversational turn is associated with the teacher or the student (cf. Chapter 2). In doctor- patient interviews the use in questions and answers is related to the interview format where the doctor asks questions and the patient provides information. In the following example from a broadcast discussion, well has the function of signal-ling the introduction of a new speaker in the debate:

(1) A: Melvyn Bragg youre President of the National Campaign for the Arts the lead signatory in the letter part of which I quoted a few moments ago

What do you thinks gone wrong B: Well before we start to talk about fi nances whichll occupy a lot of this pro-

gramme and blame whichll occupy a lot of this programme the reason why Im were here and people are watching is because most people think that the arts add something to their lives that nothing else will give them

(S1B- 022 00120014 BROADCAST DISCUSSION)

Pragmatic markers typically occur at transitions in the discourse where the hearer needs to be made aware that a new activity starts or that the speaker takes on a new role. On the metapragmatic level the hearer switches to making inferences on the basis of the interpre-tative debate frame specifying that new speakers should be introduced by a moderator in the discussion. However, many social events contain both fi xed parts which need to be interpreted according to a frame (such as the classroom frame or the telephone frame) and parts where speakers step out of their roles as teacher or student or caller or answerer in the telephone conversation. It follows that the hearer has to shift back and forth from interpreting what is said according to shared stereotypic knowledge to interpret what is said on the basis of general pragmatic rules drawing upon what is said to initiate a new utterance. Using a formulation from Silverstein we could say that there is no language use without a constant calibration between pragmatic and metapragmatic functioning (Verschueren 2000: 445). The use of contextualisation cues and interpretative frames to describe background knowledge needed for interpretation is not a new idea. According to Levinson (1979: 371), we need to consider inferential schemata tied to (derived from, if one likes) the structured properties of the activity in question. In the following chapters we will use



the term frame to refer to such inferential schemata. These are needed to explain what pragmatic markers contribute to the interpretation of the discourse.

1.3 Methodology

Corpus studies of pragmatic markers in diff erent varieties of English are still fairly new. Studies of pragmatic markers in diff erent types of activity such as courtroom examina-tions, the classroom, and other task- oriented dialogues have mostly been based on data collected by the researcher him- or herself. In such an approach, the use of observation and ethnographic fi eldwork data goes hand in hand with control of the data which is dif-fi cult to achieve by other means. Corpora, on the other hand, provide many advantages if we want to compare pragmatic markers in diff erent text types (activities and regional varieties). Luckily there now exist many new corpora and these make it possible to compare pragmatic markers both across text types and across regional and social varieties. As Mair (2009: 8) has pointed out, the rapidly growing number of publicly available corpora of English combine an increasing amount of material which sociolinguists would disregard at their peril. Below I will discuss the possibilities of using many diff erent corpora to study how pragmatic markers depend on factors such as such as text type, regional variety, and changes over time. My main resource has been the ICE- GB Corpus (the British Component of the International Corpus of English).1 The ICE- GB Corpus was collected within a project aiming at documenting varieties of English on a regional basis and is therefore also suit-able for making comparisons with other regional varieties (see also Chapter 4). Corpora make it possible to make quantitative and qualitative observations about the function of pragmatic markers, their position in the utterance and the discourse, as well as their collocations and prosody. A special advantage with the ICE- GB corpus is that it can be used as a sociolinguistic corpus. It represents a wide range of spoken and written text types and makes it possible to study pragmatic markers in many situations and with a variety of speakers in diff erent social roles. The corpus is small (one million words) in comparison with other existing corpora but has a fairly large spoken component (600,000 words) collected between 1990 and 1993. Since pragmatic markers occur mainly in speech, I have concentrated on the spoken data from the ICE- GB Corpus. However, in analysing actually and in fact I will also refer to the frequencies in the written part of the corpus. The wide coverage of text types makes it possible to show both where a particular marker occurs and where it does not occur. The largest material in the spoken corpus represents informal conversation (private dialogue) which may be explained by the fact that this is the most interesting text type for studying spoken phenomena. However pragmatic markers need to be studied in many more text types. As Lam points out, there are few studies of pragmatic markers across social situations or text types although we can expect pragmatic markers to do quite diff erent things depending on who uses them and for what purposes (Lam 2009). In the ICE- GB the general categories are, for instance, private dialogue and public dialogue. Public dialogue consists of classroom lessons, broadcast discussions, broadcast interviews, parliamentary debates, legal cross- examinations and business transactions. Not all categories are self- explanatory. Legal presentation refers for example to the pres-


INTRODUCTION 9

entation of a legal case by an attorney in court. Spontaneous commentaries consist both of sports commentaries and commentaries on public events. Categories such as broadcast discussion or classroom lesson may be considered speech events (Hymes 1972), social situations (Ochs 1996) or activity types (Levinson 1979). Activity types can diff er with regard to topic, speakers, the relationship between speakers, purpose, etc. The present study will focus on some activities represented by the texts in the ICE- GB. An advantage of the ICE- GB corpus is that it is possible to listen to the spoken recordings accompanying the texts. The prosodic contextualisation features which are of interest to the study of pragmatic markers include prosody, accent and intonation but also rhythm, pace of delivery, loudness and related phenomena such as pausing, overlaps and repetition. Both prosody and pronunciation can be assumed to be a part of the patterns associated with the speakers shared knowledge about the use of the markers. Well, for example, is pronounced both with and without a full vowel. However, the examples have not been transcribed and I have had to rely on my own prosodic analysis of the data. In order to study pragmatic markers such as and that sort of thing and or something and their formal and functional varieties in diff erent regional varieties I have used some of the ICE Corpora which are available (see Chapter 4). Recent changes in the use of prag-matic markers can be studied in the Diachronic Corpus of Present- Day Spoken English (DCPSE). The DCPSE Corpus compares data from the London- Lund Corpus (LLC) of Spoken English (mainly from the 1960s and 1970s) with similar texts from the ICE- GB Corpus. Ongoing changes and new uses of the pragmatic markers can also be studied in the conversations among London adolescents in the COLT Corpus (the Bergen Corpus of London Teenagers) (cf. Chapter 2, footnotes 4 and 5).

1.4 Linguistic theories accounting for the relationship between pragmatic markers and context

All utterances are embedded in a social, cultural and linguistic context. What makes prag-matic markers unique is that they allow context into the linguistic analysis (Verschueren 1999: 111). Not surprisingly, pragmatic markers can therefore be regarded as a testing ground for contextual theories of meaning. [I]n diff erent varieties of pragmatic theory, from Relevance Theory to more conservative Gricean theories, current work is addressed to explaining how almost vacuous or semantically general expressions can have deter-minate interpretations in particular contexts (Gumperz and Levinson 1996: 8). The situation is characterised by a tension between theories focusing on the use of pragmatic markers when people talk to each other and theories focusing on their production and understanding as a cognitive ability. According to Blakemore (2002: 5), for instance, the object of study [in pragmatics] is not discourse, but the cognitive processes underlying successful linguistic communication, and the expressions which have been labelled as discourse markers must be analysed in terms of their output to those processes. A theory of the latter kind is relevance theory. In this section I will review several theories describing how meaning is created in or interacts with the context. Integrative theories (1.4.1) describe systematic relation-ships between language and context on many diff erent levels. Relevance theory will be described in Section 1.4.2. and the theory of meaning potentials in Section 1.4.3. The



theory of meaning potentials will be further developed in Chapter 2 where I argue that the meaning of well is not fi xed but well can best be described as having a maximally rich meaning or meaning potential, part of which is activated in the situation.

1.4.1 Integrative theories

Integrative theories provide a rich description of what pragmatic markers are doing in the discourse. Starting with Schiff rins pioneering study of discourse markers (1987), the functions of pragmatic markers were analysed with regard to the structure of the discourse and more recent studies have all acknowledged the close relationship between discourse structure and pragmatic markers. Schiff rin starts out with the assumption that:

Not only does language always occur in a context but its patterns of form and function, and at surface and underlying levels are sensitive to features of that context . . . In sum I assume that language is potentially sensitive to all of the contexts in which it occurs, and even more strongly, that language refl ects those contexts because it helps to constitute them. (Schiff rin 1987: 4)

Schiff rin defi ned discourse markers operationally as sequentially dependent elements which bracket units of talk (1987: 31) and attempted to describe in a systematic way the discourse in which markers occurred (ibid. 312). A number of diff erent dimensions are distinguished making it possible to describe the multifunctionality of the markers (ibid. 24f.). Discourse is viewed as action constrained by linear sequence (e.g. what precedes and what follows), the participation framework (the relationship between speaker/hearer and between speaker/utterance). Turns are constituted through the exchange structure (the sequentially defi ned structure also known as adjacency pair). Other structures in Schiff rins model have to do with information state (who has access to knowledge and what is shared knowledge) and the ideational structure (e.g. topic relations or cohesive relations). All markers have indexical functions. Well, for instance, indexes the utterance to the speaker (and to the hearer) and to the prior or upcoming text. Schiff rins model has been important for the present work; for example, in the atten-tion to many diff erent aspects of the communication situation and the importance of indexicality and deixis to explain the role of pragmatic markers in the context. Schiff rin describes her model as integrative since it involves multiple contextual components which contribute to the overall sense of the coherence of discourse. stman does not describe his model as integrated or integrative. However, it is also an attempt to view pragmatic markers from more than one perspective (stman 1995). According to stman, communication does not operate according to Rationality principles (such as the relevance principle) but according to Common Sense prin-ciples, which need their own analytical tool box (ibid. 106). Pragmatics (in a broad sense) provides us with the parameters and potential values that linguistic elements have when they are looked at from a pragmatic perspective, for example, with regard to society, discourse, culture, human relationships (ibid. 107). This model is also compatible with the pragmatic markers being ambivalent or variable in meaning a both- and manifestation. They have values with regard to all the parameters. However, a certain value can be highlighted (foregrounded) or backgrounded in the communica-


INTRODUCTION 11

tion situation. stmans model will be further discussed in Section 1.7 and in more detail in Chapter 2 since it provides the functional typology I have used to analyse well.

1.4.2 Relevance theory

Blakemore regards relevance theoretic pragmatics as having the potential to provide a theory of utterance interpretation which is consistent with generative grammar (Blakemore 2002: 7). The contribution of relevance theory would be to explain the role of inference for utterance interpretation while generative grammar provides an account of the coding- decoding mechanism in communication. Let us consider how relevance theory can be applied to the multifunctionality of well (cf. Chapter 2, Section 2.2). Relevance theory accounts for the frustrating elusiveness of well associated with its changing purposes by referring to a communicative principle of relevance which allows the hearer to infer the intended interpretation. A pragmatic marker such as well can be viewed as a signal from the speaker allowing the hearer to make the assumptions required to recover an interpretation consistent with the guarantee of optimal relevance (every act of ostensive communication communicates a presumption of relevance). Relevance theory takes a hearer perspective and regards pragmatic markers as a signal to the hearer to take an inferential route. According to Blakemore (2002: 47)2, well could be regarded as a signal simply in the sense that it provides a green light for the hearer, a sign to go ahead with the inferential process involved in the derivation of cognitive eff ects. Thus Innes (2010), for example, refers to four diff erent functions of well (perceived needs) and suggests that the choice between them is made by the hearer on the basis of optimal relevance:

Relevance theory may be a useful framework in that it can account for wells diff erent functions within the same account; that is, the structural (cohesion and delay) and attitudinal (including face and politeness) can be gathered under the umbrella of procedural encoding and optimal relevance. (Innes 2010: 115)

However, theories focusing only or mainly on fi nding a common principle explaining how the diff erent functions of well can be inferred on the basis of contextual assumptions may not be suffi cient to explain how well is interpreted in relation to the text type or the role of the speaker in the actual speech situation.

[T]he possible eff ects arising from a number of linguistic and contextual factors on the use of discourse particles might be understated in such an approach [as relevance theory], and one of the crucial missing elements is the infl uence of text type in the interpretation of the pragmatic meanings of discourse particles. (Lam 2009: 354)

Relevance theory does not take an integrated view on how utterance meaning is achieved and is therefore less adequate to describe what is going on in talk- in- interaction. In par-ticular, it does not explain how pragmatic markers can change over time or in diff erent text types. We need also to consider politeness principles and appropriateness conditions



which are less general than the relevance principle but specifi c to a particular culture or society, region, social situation, historical period, etc.

1.4.3 Pragmatic markers and meaning potentials

It can be claimed that pragmatic markers do not have a fi xed meaning but a meaning potential. The theory of meaning potential can cope with the meaning representation of lexical items which have no strictly delimited meanings but develop meanings in situ-ated use (Norn and Linell 2007). According to Norn and Linell (2007: 387), the basic assumptions of a theory of meaning potentials are that the linguistic resources provide language users with semantic resources to understand, say and mean specifi c things in particular usage events, and that this always involves an interplay with contextual factors. The theory of meaning potentials provides a rich meaning description of pragmatic markers where the interaction with the context is important. According to Norn and Linell (2007: 390), (a) theory of meaning potentials assumes that parts of a words meaning are evoked, activated or materialised, foregrounded or backgrounded, in dif-ferent ways in the diff erent contexts, in which it is exploited. Fischer, for example, found that diff erent meanings of (the meaning potential of) okay were activated depending on whether the situation was human- to- computer communication or communication between humans. She found that speakers did not use okay to take their turn or play for time although these functions were frequent in communication with another human being (Fischer 2006: 441). The theory of meaning potentials describes the relationship between meaning and context in a diff erent way from relevance theory: the context selects the meaning of a pragmatic marker whereas in relevance theory a pragmatic marker looks for a context which is compatible with communicative principles. If meaning is selected in the com-munication situation we need a description of the contextual factors which interact with the meaning potential of the lexical item. As will be shown in Section 1.5, the relevant or selected context can include text type and speakers and related features such as speaker identities, professions, etc. which are indexed in the communicative situation along with discourse and pragmatic functions. In addition, formal features (the linguistic co- text) can be regarded as being part of the context (or usage domain). The meaning potential approach can also be compared with the rich description of the formal, functional and contextual aspects of pragmatic markers in a construction gram-matical framework (Fried and stman 2005; stman 2006). Like constructions, meaning potentials represent part of the language users knowledge of language. It seems rather that meaning potentials are part of the actors knowledge of language, and they are used in the negotiation of situated interpretations (Linell 1998: 82 quoted from Keevallik 2003: 32).3 In particular, we can look upon meaning potentials as an economical way of storing the speakers knowledge about language. We also need to consider how the information is stored. The meaning potential of pragmatic markers is not an unordered set of mean-ings but the potential must have some internal structure. The theory is compatible with the polysemy of pragmatic markers, that is, the idea that pragmatic markers have one or several core meanings from which new functions can be created in the interaction. It has for instance been suggested that the meanings of well can be organised into two broad spaces: that of modal functions and that of structural functions (Cuenca 2008: 1388). In


INTRODUCTION 13

Cuencas analysis well has a radial structure or semantic networks with more peripheral meanings organised around one of the core aspects. Meaning potentials organise (only) the conventionalised meanings of pragmatic markers. The conventionalised meanings are important because they can be the basis for classifying the pragmatic markers. They are distinguished from less conventionalised (or ad hoc) meanings created in the communica-tion situation. The theory of meaning potentials is compatible with what we know about how mean-ings can be negotiated, modifi ed or changed diachronically (both in a short and a long time perspective). It can also explain what we know about how pragmatic markers are acquired. Children and language- learning novices fi rst acquire pragmatic markers in prototypical face- to- face situations and then start using them in more situations and new contexts when they are socialised into new ways of life. It has, for instance, been shown that English- speaking children acquire the potential for systematic sociolinguistic varia-tion together with pragmatic markers from an early age (Andersen et al. 1999: 1349). Meaning potentials are potentially creative and make possible all the usages and interpretations of the word or construction that language users fi nd reasonably correct, or plainly reasonable in the actual situations of use (Norn and Linell 2007: 389). They can therefore explain how markers can be used in text- type specifi c or innovative ways in the communication situation. Greasley (1994), for example, investigated the use of well in television commentaries on the game of snooker and found that well was produced under certain contextual conditions such as a surprisingly good shot. Since well is used we can assume that speakers experienced the situation as similar in some ways to prototypical situations where well is needed and therefore extended it to new situations. To sum up, pragmatic markers have a variable and context- bound character. Their meaning therefore cannot be described in the same way as lexical elements which have a more stable lexical meaning. The theory of meaning potentials accounts for the fact that pragmatic markers get their meaning through dynamic sense- making in local, situated contexts (Norn and Linell 2007). Finally, we should note that the theory of meaning potentials is appropriate to describe the diff erent uses of pragmatic markers in new text types in the corpora. We can use it to explain both established or conventionalised mean-ings and innovative or ad hoc meanings characteristic of special activities.

1.5 Pragmatic markers and the context

Pragmatic markers get their meaning in interaction with the context. It is therefore important to defi ne what is meant by context. Context needs to be taken in a broad sense. In the anthropological literature we fi nd a rich description of the social situation (social event, activity type) which is relevant for the analysis of pragmatic markers. The descrip-tion of the social dimensions below is based on Ochs (1996). Ochs suggests that linguistic elements can reach beyond the utterance and index elements such as the speaker, speaker identities and activity and especially the speakers (epistemic and aff ective) stance. The notion explaining the association between pragmatic markers and the context is assumed here to be indexicality. Indexicality has an entrenched position in the deictic systems associated with time, space and person in diff erent languages. However, studies of deixis in diff erent languages generally focus on structures or elements indexing a single situational dimension only. As a result, we know a great deal about adverbs and



pronominal forms indexing dimensions such as the time and place of the activity or the speaker/hearer framework. In the sentence The book is here, here indexically shifts its interpretation depending on the place where the communication takes place. In comparison with other deictic elements, the pragmatic markers have a rich social meaning since they are used to indexically refer to a number of speech act features such as the speaker and the hearer, social identities and the speech event (social activity) itself.

While a number of studies of language use dwell on the relation of linguistic forms to only one situational dimension and ignore social dimensions socio- culturally linked to that dimension, other studies . . . consider a range of situ-ational dimensions socio- culturally entailed by a set of linguistic forms. (Ochs 1996: 418)

Ochs identifi es the following socio- cultural dimensions of the communicative situation (other than time or place):

Social identity encompasses all dimensions of social personae, including roles (e.g. speaker, overhearer, doctor, teacher), relationships (e.g. kinship, friend-ship), group identity (gender, generation, class, ethnic membership) and rank (employer and employee).Social act refers to a socially recognised goal- directed behavior (e.g. a request, an off er)Activity refers to a sequence of at least two social acts, e.g. disputing, interviewing4Aff ective stance refers to a mood, attitude, feeling, and disposition, as well as degrees of emotional intensityEpistemic stance refers to knowledge or belief vis- - vis some focus of concern, including degrees of certainty of knowledge(quoted from Ochs 1996: 410 with some abbreviations)

An important dimension of the communicative situation is stance. Fischer (2006: 445) observes for example: Thus discourse particles [pragmatic markers] make the human interlocutors (with their hopes, fears, desires, and imperfections) part of the situation. However it is not immediately obvious that pragmatic markers can express epistemic and aff ective stance. Less recognized is the fact that, in many languages, aff ective and epis-temic stance is encoded at many levels of linguistic structure (Ochs 1996: 412). Ochss examples are sentential adverbs (disjuncts) and constructions with epistemic verbs (I think) which are closely associated with speakers and their feelings. Ochs does not discuss pragmatic markers but it can be argued that many of them have an important stance- marking function and that they can express a large number of stances depending on the text type. A speaker confronts numerous possibilities in framing and executing her speech and take up a stance depending on the preceding context and how they want to appear (Mertz and Yovel 2003: 6). Epistemic stance can be associated with uncertainty and politeness. In other cases, an element of disagreement or fi ghting is conveyed by the lexical element. The component of fi ghting is illustrated by the following example from Ochs (1996: 423), borrowed from M. H. Goodwin (1990) in which Ruby uses I know not (only) to


INTRODUCTION 15

express certainty but to construct a challenge to Staceys assumption that its a free world is news to him:

Stacey: Fight yourself.Ruby: Well you make me fi ght myself.Stacey: I cant make you. Cuz its a free world.Ruby: I know its a free world.

Even agreeing with the preceding speaker can carry with it an element of fi ghting or divergent opinion as shown by this example. Pragmatic markers are contextual resources enabling speakers to express a number of new stances in the interaction. In particular, they can have a rhetorical function allowing the speaker to take up a stance of alignment or disalignment to the hearer or to what is said. It is easy to overlook the fact that pragmatic markers can also be used to express authority and power (an authoritative or knowledgeable stance) associated with a par-ticular social or professional identity. As shown by Ochs, there is link between stance and social identity conveyed by cultural expectations. For example the speaker may display an authoritative stance in order to build up an identity as a medical doctor. However pragmatic markers are fl exible and can be used with diff erent stances and exhibit diff er-ent speaker statuses depending on the social activity and the speaker role. Well can, for instance, be used by a prosecutor in a court examination to display an authoritative stance. If well is used by the witness, on the other hand, it is associated with a deferential attitude. In such examples (which will be further illustrated in Chapter 2), well does not merely point to a social identity but helps to constitute or construct that identity (Ochs 1996: 424). Aff ective stance is also important with pragmatic markers. Well can, for instance, signal disappointment, resignation or reluctance. As will be further illustrated in the empirical studies, pragmatic markers have a rich situational meaning which is exploited in the communication situation also for the speakers manipulative purposes. If one does not understand what roles pragmatic markers have in communication and how they vary their functions situationally, this can have serious consequences for example in court proceedings (Hale 1999). As Hale points out (cf. Chapter 2: Section 2.2), the translators (interpreters) often mistranslated well (in the prosecutors questions) or omitted it from their translation which resulted in misunderstandings. Let us note fi nally that even though there are many cues to the intended interpretation of pragmatic markers in the context we should not underestimate the problems of convey-ing or interpreting the situational meaning. The fl exible and fuzzy meaning of pragmatic markers places a burden on both the speaker and the hearer in the communication situ-ation. The speaker must have access to a large number of linguistic and extra- linguistic conventions in order to use pragmatic markers. Hearers use a number of clues such as the lexical meaning, grammatical features (such as position or collocation), gestures and prosody to interpret the situation meaning of pragmatic markers. Even so, misunderstandings are common. According to Ochs (1996: 413):

[i]t is important to stress at this point that the assignment of situational mean-ings is a complex, interactionally accomplished process. Interlocutors have



available to them a reserve of linguistic structures some grammatical, others discursive that are conventionally associated with particular situational dimen-sions. Interlocutors may use these structures to index a particular identity, aff ect, or other situational meaning; however, others co- present may not neces-sarily assign the same meaning.

1.6 Formal features of pragmatic markers

As argued in Section 1.4.3, pragmatic markers can be regarded as lexical resources associated with a meaning potential. They also have formal properties which are part of the description of their usage. These can be of diff erent kinds: The more one studies discourse particles, the more conventions one discovers with regard to their use. The features are syntactic (e.g. position, prosodic, lexical (e.g. collocations) and stylistic (e.g. text type) (Aijmer 2002: 28). However, few studies look at the formal features in detail, probably because it is diffi cult to establish a clear link between form and what pragmatic markers are doing in communication. For example, pragmatic markers do not occur anywhere in the utterance or the turn but there are rules for their placement which also have to do with their function. In addi-tion, sequential information is important. Pragmatic markers can point both backwards or forwards to project a new turn. They can also have a global textual function and point forwards to a narrative, argument or description. They occur in diff erent linguistic con-texts and can have diff erent functions depending on whether they occur in questions or answers. Prosodic (and other formal) features have been discussed as clues to the function of pragmatic markers and as criteria for their status as pragmatic markers. For example, in the area of speech recognition and understanding, pragmatic markers have been regarded as cue phrases which together with prosodic and grammatical uses (such as position in the utterance) constitute important information for understanding how the utterance is segmented or to disambiguate its diff erent meanings or functions (Hirschberg and Litman 1993; Horne et al. 2001). The relationship between form, function and prosody is even more complicated if a pragmatic marker is multifunctional. At least in some cases we fi nd interesting correla-tions between form and function. Ferraras (1997) study of anyway is of particular inter-est in this respect. Ferrara established three diff erent meanings or subtypes of anyway which could be distinguished on the basis of syntactic position and prosody. Ferraras analysis is compatible with the use of prosody as a cue (along with other features) enabling the hearer to identify the function of a particular pragmatic marker in its context. However, not everything is conventionalised: [. . .] prosodic choices [. . .] are motivated in part by convention, but largely by the often confl icting demands and constraints of the semantic, pragmatic and discoursal functions that discourse markers fulfi l (Wichmann et al. 2010: 150). It is obvious that formal features do not have a simple relationship to function. Thus, for example, a pragmatic marker in initial position (and a certain prosody) can be a cue to a discourse function (e.g. the use of a pragmatic marker to introduce a new topic). However, the same function may be fulfi lled by other pragmatic markers which are placed initially and have a certain prosody (what Bazzanella 2006: 454 refers to as their (partial) interreplaceability). Fischer (2006: 443) gives examples of how


INTRODUCTION 17

okay, hmm, oh or well can all function to signal successful perception, understanding and topic continuity although they do it in diff erent ways. Both interreplaceability and the examples where it fails must be accounted for. The hypothesis here is that many diff erent factors are involved in describing why a certain pragmatic marker is used. These factors include for example the speakers knowledge that a pragmatic marker is associated with a certain text type, is used by a speaker in power, etc. In Chapter 3 on actually and in fact, it will be shown that even pragmatic markers which sometimes collide in a par-ticular function are associated with diff erent usage conventions and can be experienced diff erently.

1.7 Functional features of pragmatic markers

The distinction between two or more basic functions of pragmatic markers is made in many functional models. A controversial issue is however how many parameters (domains, basic functions) we should distinguish. Fischer (2006: 430) considers a wide range of functions that pragmatic markers (Fischers discourse particles) can have:

Although scholars disagree much about the functions of discourse particles, there is a range of functions that is commonly, and often cumulatively, attrib-uted to them. This spectrum includes functions with respect to the turn- taking system, the indication of discourse relations, discourse structuring, the regula-tion of interpersonal relationships, speech management, or politeness.

Schiff rins integrative proposal involving fi ve diff erent discourse planes was dis-cussed in Section 1.4.1. However, many proposals distinguish fewer dimensions. Brinton (1996 and 2008) distinguishes two rather than three major functions. Pragmatic markers serve a textual function related to the structuring of discourse as a text and an interper-sonal function which relates to the expression of speaker attitude and to the organization of the social exchange (Brinton 2008). stman distinguishes three parameters in accordance with which communication takes place (1995: 104) and with reference to which research into implicit pragmatics is doable: For me, then, discourse marking or discourse organizing is one major func-tion that pragmatic particles have: a good candidate for another major function would be interaction- signalling; and yet another would be attitude/involvement signalling (stman 1995: 99). We will return to stmans functional typology in Chapter 2, where it is used to sub- classify the functions of well.

1.8 Summary and conclusion

In this introductory chapter I have emphasised the need for a new study of pragmatic markers which focuses on their uses in diff erent social and regional varieties. It will be shown that pragmatic markers function very diff erently depending on text type, activity, speaker roles, regional variety, etc. I have also discussed the type of theoretical frame-work we need to describe pragmatic markers and some basic assumptions underlying my approach.



Pragmatic markers have been characterised as metalinguistic indicators which attend closely to what is happening in the communication situation and remark explicitly on aspects of the ongoing speech event. They are used to manage the speakers moment- to- moment progression from one turn to another. This analysis is not suffi cient to describe how they are used in diff erent situations. It has been argued that they can also be used as contextualisation cues indexing a frame for the interpretation of a pragmatic marker such as well when it has specialised uses (discourse tasks). Speakers also make use of stance to take up diff erent positions to the hearer or to the text in the interaction. Traditionally it has been assumed that the context referred to is the linguistic environ-ment in which the pragmatic markers are used. However, pragmatic markers also have a number of innovative or ad hoc meanings depending on the speaker or the social activity in which they occur. Pragmatic markers have indexically rich sociolinguistic or situ-ational meanings. To explain these we may have to go beyond the constraints imposed by the need to be coherent, polite or to express emotions and feelings. They can be indexed to elements in the air such as the speaker, the hearer, (epistemic and aff ective) stance, social identities and a social activity. The social or situational meaning can be indirect, for example when a pragmatic marker is used by the speaker to construct a certain social identity or professional identity (for example, as a teacher or a doctor). In such cases, the explanation of the meaning of a pragmatic marker involves cultural expectations, for example, about power status in society. When we consider how pragmatic markers are used in diff erent text types it becomes diffi cult to maintain that they have a fi xed meaning. It has been argued that pragmatic markers construct meaning potentials based on their uses in many diff erent varieties. The meaning potential is in constant fl ux and is used creatively in new situations. In the actual communication situation the meaning potential is used selectively and perhaps with ad hoc or innovative functions. Pragmatic markers can be looked upon as combinations of formal and functional features and descriptions of the contexts in which they are used. The formal features include, for example, pronunciation, pausing, positioning, linguistic context, collocation. However, there is not a simple relationship between form and function as will become apparent in the empirical case studies in the following chapters.

Notes

1. See the website http://www.ucl.ac.uk/english- usage/ice# (accessed 4 October 2012).

2. Well has also been described in a relevance- theoretical approach by Jucker (1993), de Klerk (2005), and Innes (2010).

3. Keevalliks orientation is, however, interactional linguistics and Conversation Analysis. Interactional linguistics combines a constructional approach and Conversation Analysis (see Keevallik 2003). It has in common with the meaning potential approach that it favours a rich representation of the meaning of lexical items. According to Deppermann (2006: 59), [i]t is in fact one of the central tenets of interactional lin-guistics to reconstruct the holistic interplay of syntax, semantics, prosody, pragmatics, multimodality, and sequential aspects of grammatical structures in conversation and to


INTRODUCTION 19

account for the contingencies of the empirically given token in all their informational richness (quoted from Imo 2009: 4; Imos translation).

4. Other terms which are used in similar ways are activity type (Levinson 1979) and com-municative activity type (Linell 2009).


2The Pragmatic Marker Well

Well, the little words are often the most interesting. (Innes 2010: 95)

2.1 Introduction

Well is interesting not least because of its frequency in conversation:

A: Why dont you uhm replace one of the back doors here and use the pane from C: Well theyre rotten now Dad so one day youll have to A: Well Im just putting that off for as long as possible C: Well if I ever have to replace a back door I shall do so (ICE- GB)

The speakers in this little dialogue preface their contributions to the conversation with well. Well is not a part of the propositional content but its function is to enable the hearer to interpret how what is said should be understood. Well as a turn- initiator is used with several diff erent functions. In the example above, well is, for example, used by C (the son) to signal that he is unwilling to replace one of the back doors because it is rotten. The father (A) uses well in a defensive way (Im putting it off for as long as I can). In the next turn the son uses well to indicate reluctant agreement (If I replace a back door I will use a pane to do so). The examples of well have in common that they make the speaker come alive by expressing his emotions and attitudes. Without well the interaction would sound abrupt or impoverished because of the absence of the speaker attitude. Not surprisingly, well has received more attention than any other English pragmatic marker (Schourup 2001: 1026). It has functioned as the testing ground for diff erent theoretical approaches to the study of pragmatic markers and it has been investigated by means of several methods. However, existing studies on well show little agreement on controversial questions such as whether well has meaning, the number of functions it has and what these functions are. There are also neglected topics, for example how the func-tion of well is allied to speaker roles and identities or to activities and text types. Well has the potential to mean diff erent things depending on the context. Does well


THE PRAGMATIC MARKER WELL 21

add anything to the interpretation of the context? If this is the case what meaning or meanings does it have? In the past well has mostly been studied in conversation. Another aim of this chapter is to extend the study of well to diff erent text types and situations to get a better picture of its formal and functional properties. The following research questions will be asked: In what text types is well frequent and why? What formal properties does well have? Where is it infrequent? What functions does it have in diff erent text types? Who uses well? How is it used to establish a social identity and role? The analysis of the functions of well in diff erent text types can also contribute to the analysis of pragmatic markers generally by drawing attention to the situational or contex-tual variables motivating the choice and function of well. Such variables may depend on the medium (the diff erence between face- to- face and telephone conversation), the diff er-ence between monologue and dialogue, the presence of a person in control of turn- taking and the topic- management, etc. Well has earlier been investigated on the basis of the London- Lund Corpus by Svartvik (1980). However, Svartvik analysed well only in a small sample of conversational texts and it is now possible to use more recent corpora than the London- Lund Corpus. The more recent ICE- GB Corpus (the British Component of the International Corpus of English) has therefore been chosen for the study of well here (see Chapter 1). The outline of this chapter is as follows. Section 2.2 reviews previous analyses of well focusing on studies of the marker in diff erent text types. The distribution of well over diff erent text types in the ICE- GB is presented in Section 2.3. The formal properties of well are discussed in Sections 2.4 and 2.5 (collocation). In Section 2.6 it is argued that well has a meaning potential rather than a fi xed meaning. Section 2.7 introduces the functional typology and the parameters used for the classifi cation. The functions are discussed under the headings coherence (Section 2.8), involvement (Section 2.9) and politeness (2.10). In Sections 2.112.13 well is discussed in some selected text types and situations. Section 2.14 contains the conclusion.

2.2 Previous studies of well

It is hardly possible to give an overview of the vast literature on well, particularly as research is now extending into new areas (cross- linguistic studies, well in learner lan-guage).1 Well has been the subject of many theoretical approaches, for example, the coherence- based theory of Schiff rin (1987), relevance theory (Jucker 1993; Blakemore 2002; de Klerk 2005), dialogue game theory (Carlson 1984) and politeness theory (Watts 1989). These theories explain the uses of well with regard to a single principle (such as relevance, or politeness) or postulate a core meaning and formulate a rule from which other functions can be derived. Carlson, for instance, attempts to make a simple rule for well work in a wide variety of cases (Carlson 1984: 2) in a theory of dialogue games. In one of the most infl uential studies of well Schourup argues for instance that well (in present- day English) can be regarded as a gestural interjection indicating a relationship between a mental state or attitude and a vocal portrayal of that state (2001: 1050). Because it is also a lexical item (as in he speaks well) it diff ers from interjections such as oh or ah which have no lexical correspondence2. A similar interjectional analysis is represented in other work and is compatible with the analysis of pragmatic markers as



metalinguistic indicators tracing the speakers progression through the discourse (cf. Chapter 1, Section 1.2.1). According to Biber et al., the speaker is neither certain nor uncertain but is thinking about things, collecting his or her thoughts when using well (1999: 1086). Compare also Carter and McCarthy (2006: 152) who, however, regard well as a discourse marker: A main function of well as a discourse marker is to indicate that the speaker is thinking about things. This understanding of well is also close to Carlsons defi nition of well as deliberative, which he defi nes as weighing alternatives in terms of acceptability in order to arrive at a decision (1984: 86). In other words, well signals an ongoing mental process in which the speaker deduces or infers something on the basis of considering the alternatives in a particular situation.3 However the analyses proposed by Schourup or Carlson are characterised by meaning minimalism: well has a single core meaning from which new meanings can be derived. This analysis will be contrasted below (Section 2.6) with a richer meaning theory making it possible to distinguish between core aspects (sub- senses, inferences, connotations, indexical meanings). Another problem is that multifunctionality may be understood too narrowly because only conversational data is considered. For example, neither Schourup nor Carlson discuss well in diff erent social contexts where we can assume that well can signal many diff erent things. Well has a large number of functions depending on the social situation (activity type), text type and speaker role in addition to the linguistic context or general features in the communication situation. A preferable starting- point is therefore to study formal and functional properties in diff erent text types without considering how the functions fi t in with a description of what well means. Well has a number of specialised functions derived from its association with a particular text type and situation (the type of activity, who the speaker is, etc). Taking a hint from studies of well in other text types than conversation we can assume that many functions of well are not only dependent on the lexical meaning of well and the linguistic context but depend on cultural conventions or norms and social dimensions of the speech situation. For natural reasons, past research has tended to focus on the give- and- take of everyday talk (Norrick 2001: 851) and the functional categories proposed are those associated with conversation. However, we are now also beginning to see a number of analyses of well in very diff erent types of text (Hale 1999; Fuller 2003; Greasley 1994; Norrick 2001; Innes 2010). They raise questions about the adequacy of earlier studies of the functions of well based on conversational data only. Below I will review some works which are of interest for the present study of well because they give evidence for the rich meaning potential of well, and in particular because they show that we need to go beyond its discourse- structuring functions. Green (2006) studied the importance of well (and other pragmatic markers) for the creation of a question- answer system which could be used for natural language process-ing. She was interested in well because it is typically attitudinal unlike markers such as now, okay, and and but:

Functionally, two basic kinds of discourse particles can be distinguished; both are conversational. Attitudinal discourse markers, such as well, uh, like, gosh, oh, OK, I mean and yknow indicate something about how the speaker feels about what is being said [. . .] (Green 2006: 118)



On the other hand, there are structural discourse markers like sentence- initial uses of now, OK, and, but which speakers use to indicate a structural boundary in the discourse. (Ibid. 119)

Greens analysis is interesting both because of the experimental approach and because she pointed to the association between well and the attitude of the speaker which she explained with reference to face- to- face contact with another speaker. Initial experimen-tal work had shown that although the interface to the program was given a human voice it still did not sound suffi ciently natural and friendly. Greens research team therefore chose to work with discourse particles (pragmatic markers) because they interact minimally with the syntax of the sentence they are attached to, while conveying a lot of information about the state of mind of the speaker (Green 2006: 121). The intended result was said to be the illusion of a mind behind the monitor screen that thinks and cares about the user whom it addresses (Green, ibid.). The reasons for using well also include the linguistic context according to Green. Well was therefore studied in diff erent contexts to fi nd out how well adjusts to diff erent conversational demands (e.g. whether the sentence is a ques-tion or a declarative sentence, the eff ect of position, etc.). The experimental design of Greens study pinpointed the important link between the speakers emotions and attitudes and the use of well in conversation. However, the distinction between attitudinal and structural is not so clear- cut as suggested by Green. Well can, for instance, be both attitudinal and structural with functions such as topic shift or indicating a boundary in the discourse. In some text types, as we shall see below, the structural function of well is even predominant. A further point is that well is not only friendly but it also conveys attitudes such as authority and power as well as rudeness in contexts characterised by unequal speaker- hearer relationships. Greens study is interesting also for the present analysis of well. Green only studied spontaneous conversation where well was typically associated with friendliness. However, depending on the situation and text type well can convey many diff erent attitudes includ-ing hesitation or uncertainty but also feelings such as disappointment or reluctance. Indirectly well can be used to construct a certain identity, personality or social persona. The association with attitudes is an important feature of many analyses of well foreshad-owed in the discussion of pragmatic markers and indexicality in Chapter 1. Of particular interest for this study is that we are now beginning to see a number of sociopragmatic studies of pragmatic markers, that is, studies paying attention to their functions and variability in diff erent social contexts. These studies suggest that well can express diff erent attitudes and functions in diff erent text types or activities. These studies also point to the importance of well to express status or stance relating to personality and professional identity. Hale investigated interpreters treatment of well (and other markers) in the courtroom. According to Hale (1999: 1350), pragmatic markers such as well add tone and force to the utterance. Well (and see) was for instance shown to be frequent in lawyers questions in cross- examinations as an assertive device, indicative of superior authority (ibid. 59). The importance of well appeared when well was not rendered by the interpreter in the court examinations or when it was mistranslated. Omitting well is always at a cost since the marker adds information about the speakers attitudes, which is important in order to understand the speakers intentions and to avoid misunderstanding. When well was



not translated in the court proceedings the style and manner of the lawyer and defendant were not expressed. Moreover, interpreters frequently chose a translation that altered the illocutionary force of the utterance and therefore aff ected the witnesses experiences of what went on during the proceedings. Hales study suggests that well has a meaning which goes beyond the discourse context. However, the function has to do with emotion and stance rather than with the discourse structure. Another courtroom study is Innes (2010). Innes discussed the use of well in jury trials in New Zealand courtrooms. The study showed that the use of well depends on the users roles and the social goals of the activity. Well was for instance used in challenging ques-tions in the cross- examination by the lawyer and with a hedging function in the defend-ants answer. Norricks study of oral narratives is of particular interest when we study pragmatic markers in diff erent text types since it shows that the functions of well cannot be dis-tinguished from aspects of the ongoing activity where it is used. According to Norrick (2001: 849), the functions of well in oral narrative can be related to the conventions for story- telling and its highly coded sequentiality. In Norricks words, well acts as a specifi cally narrative DM [discourse marker] keyed on expectations about the organiza-tion of stories (ibid. 857). Well can for instance lead into beginnings and endings of oral narratives and mark a return to the main theme of the story (the stage- marker or staging function). Mller (2004) who analysed oral narratives based on a silent Chaplin movie came up with related categories such as move to story and introducing the next scene.4 Besides narratives and courtroom proceedings, classroom discourse has attracted con-siderable attention because of its clear structure. Sinclair and Coulthard (1975), in their pioneering research, have shown how classroom interaction is regulated by the activities in the classroom and the specifi c roles of teachers and students. Well has an important function to signal diff erent stages of the interaction. Well could, for example, occur as a framing move in a conversational exchange in the opening of the classroom lesson or to mark the transition from one stage of the discourse to another. Sinclair and Coulthards work has also been an inspiration for a study of (Spanish) pragmatic markers in the class-room by De Fina (1997). De Fina (1997) was concerned with describing the functions of Spanish bien (a prag-matic marker with functions overlapping with well) in classroom discourse. In her study bien was seen to have two classroom specifi c functions (both of which were associated with the teacher and not with the students). Transitional bien was for instance used by teach-ers to signal changes in classroom activities, while evaluative bien functioned as a feedback move. Both functions could be associated with the degree of control of the teacher in the interaction. The transitional use of bien occurred at transitions and boundaries in the discourse, for example, to indicate that one phase of the classroom activity is fi nished and there is a change to a new situation with a diff erent alignment of speakers. De Fina (1997) also suggested that Gumperzs term contextualization cue could be used to mark the role of the transitional bien to change or redefi ne the context (cf. Chapter 1, Section 1.2.1). Fuller (2003) discussed pragmatic markers in interviews and showed that well was used more frequently (but not exclusively) by the interviewer than by the interviewee and that it was found in diff erent patterns:



The DMs [discourse markers] oh and well are used relatively infrequently in the interviews by the interviewees, but at high rates in these same interactions by the interviewers. Because the interviewees are playing the role of the speaker more than of the listener, they are not expected to off er as many responses to the contributions of the interviewer as they would in a symmetrical conversation. The interviewer, however, is primarily a listener and thus uses these response signals more frequently. (Fuller 2003: 43)

Documents

understanding pragmatic markers.pdf